Why can hardware work without software

Find and fix hardware problems

David Wolski

Hardware problems do not provide uniform error patterns and tend to arise spontaneously at inconvenient moments. A systematic approach promises a quick end to the hardware horror.

EnlargeHow to analyze hardware errors and eliminate them in next to no time.
© © Nomad_Soul - Fotolia.com

Problems with aging and defective hardware do not have to show up in the event of a total computer failure, but rather manifest themselves in a very subtle way. If computer problems accumulate, a reproducible error simply cannot be eliminated by changing the configuration of the operating system, or a certain hardware component can no longer be persuaded to work properly, precise and systematic troubleshooting is required. A solid, well-founded strategy can help avoid both time-consuming and unnecessary steps in the first place.

The following article is a guide on how to isolate and diagnose hardware errors without extensive tools. This article is all about the analysis and the cause of the error: Find out whether the hardware and not perhaps the software is responsible for the trouble with the computer.

1. Bios and Uefi

1.1 Replace the CMOS battery on the motherboard

The PC forgets the time and BIOS settings. Error message at system start: "Mismatch CMOS" or "CMOS Read Error".

Diagnosis and repair: So that the settings in the BIOS (Basic Input / Output System) or Uefi, i.e. the firmware of the main board, are retained even when the power supply is switched off, the board has its own button cell battery. This serves as a buffer for the otherwise volatile CMOS memory of the firmware. Typically, the battery will last three to five years. If you have caught a Monday product, the battery can run out much earlier. Features of an empty button cell are reset BIOS / Unicef ​​settings, an unreliable system clock and the message “Mismatch CMOS” or “CMOS Read Error” when the system is started.

EnlargeThe CMOS battery is located on the main board: Since the button cells only last five years, all PC users with older computers will at some point be confronted with a forgetful BIOS / Uefi.

In the event that you receive one of the above messages, you should switch off your PC, disconnect it from the power supply, open the housing and locate the location of the button cell on the motherboard. With very few exceptions, the battery is a lithium button cell of the type CR2032 with 3 volts and 225 to 230 mAh. Since this type of battery is also used in watches and car keys, these button cells (from 1.50 euros) are available in almost every supermarket or at three at night at the petrol station. You do not need any special tools for the change; a flat screwdriver or phase tester is usually sufficient to remove the old battery.

If you have several button cells in your fund, but you are not sure whether they are still fresh enough due to the lack of a battery tester, a simple trick can help: briefly connect a green, blue or white light-emitting diode with a suitable forward voltage of around 3 V. If the LED glows when the polarity is correct, the button cell is OK. The internal resistance of the battery is the series resistance here. Normally, however, you should only connect LEDs with a series resistor, otherwise the LED will not last long.

Uefi: This is how the successor to Bios works

1.2 What to do when the BIOS beeps

Beeps over the PC loudspeaker when starting the computer clearly indicate a hardware problem. It could be a defect or a harmless contact problem with a component.

Fault finding: Unfortunately, the BIOS manufacturers have not yet been able to agree on standards. Every BIOS beeps - the most common ones come from AMI, Award and Phoenix - its error codes beep significantly differently. The only thing that helps here is a look at the motherboard documentation or a search on a website such as www.bios-info.de. The "error signals" of all common BIOS variants are explained in detail here. The most common beep codes cause graphics cards and RAM modules, which can often be corrected by removing and reinserting them.

2. Overheating: Problems with waste heat

Although computers have self-protection mechanisms that protect a CPU and the motherboard from death by heat and switch off the system if necessary, problems with overheated systems are usually subtle. The errors range from occasional program and game crashes to blue screens and sudden restarts.

Fault finding: Even if current desktop PCs with energy-efficient processors and RAM modules rarely suffer from heat build-up, the problem of inadequately cooled components arises in home-made PCs, with housings that are too narrow or simply with failed fans. A common feature of the initially arbitrary errors: They only occur after a certain operating time or when the computer is very busy. If the CPU suffers from poor cooling, the system also slows down noticeably and reacts increasingly sluggishly.

EnlargeThis older PC was no longer stable and Speedfan showed excessive temperatures of the CPU and motherboard. The cause is self-explanatory after a look into the housing.

1. Watch the fans with the case open. Too much dust and blocked ventilation slots in narrow mini-PCs have a huge impact on cooling.

2. In the BIOS or in the Uefi settings, a menu item shows the temperatures of the processor and motherboard, but almost in idle without actual load. The freeware program Speedfan delivers realistic values ​​during operation. After the installation and the first call, Speedfan needs about a minute to recognize all sensors. The tool then displays the temperatures under "Readings" in the right column. "GPU" corresponds to the graphics chip, "Temp1 / 2" is the system temperature and from "Core 0" the CPU cores are displayed.

3. If the operating temperature is still in the normal range below 70 degrees, then carry out a test with a targeted high system load and keep an eye on the temperatures. The Linx benchmark program is recommended for testing the CPU. You give the GPU something to do with the Furmark graphics benchmark. Cancel the test if the CPU temperature goes well above 70 degrees Celsius, as this is the maximum operating temperature of the CPU specified by Intel and AMD. If Windows hangs during the test, it is mostly caused by heat problems in the CPU or problems with the memory modules (RAM).